On the meaning of "plasticity" in neuroscience and mental health research and its relation to the action of psychedelic therapy

Chapter 1

On the meaning of "plasticity" in neuroscience and mental health research and its relation to the action of psychedelic therapy

First-line dictionary definitions of "plasticity" describe it as the ability (of a thing) to be shaped or molded; we call this "plasticity proper". In neuroscience, however, "neural plasticity" or "neuroplasticity" is defined as an induced change in brain function or structure. These two plasticities are not the same. Several popular biomarkers of neuroplasticity index part of an evoked process, often linked to learning. Paradoxically, these processes often bias phenotypic canalization, the opposite of phenotypic plasticity. Recognizing this paradox puts into question the custom of using "plasticity" as a shorthand for all neuroplastic phenomena. This custom risks an overgeneralization fallacy and the conflation of assays of evoked brain changes with plasticity proper. We also question extrapolating from any biomarker of neuroplasticity to improved mental health, whether it be aligned with plasticity proper or not. The relationship between the two is, logically, context dependent. As a resolution, we propose a new construct, "mediational and recalibrative plasticity" and show how markers of it can describe and predict examples of phenotypic plasticity, where those phenotypes are states and traits relevant to mental health. In contrast to many markers of neuroplasticity, MR-P is aligned with plasticity proper.

One Introduction

The most basic dictionary definition of plasticity states it to be: "the quality or state of being plastic," especially the "capacity for being molded or altered". In this article, we refer to this orthodox plasticity as "plasticity proper". Plasticity proper is domain-general-it is a basic computational property that can be applied to any phenomenon, whether that be a property of brain function or anatomy-in the domains of neuroscience and biology, a psychological state or trait-in the domain of psychology and psychiatry, or even the malleability of a substance, such as a metal, in the domains of physics and metallurgy.

It is common to see first-line dictionary definitions of plasticity followed by an overtly biological definition, reading something such as: "the adaptability of a phenotype (i.e., a property of an organism) to changes in its environment". This definition is aligned with plasticity proper. The effect of increasing the plasticity/changeability of a phenomenon, for example, a psychological phenotype such as a state or trait relevant to mental health, should, by implication, render that phenotype more responsive to, and potentially adaptable to, its environment. The biology-specific definition of plasticity is sometimes referred to as "phenotypic plasticity". Conveniently, it is used consistently in psychology. It is intentional that phenotypic plasticity, psychological plasticity, and plasticity proper share the acronym "PP" as each "plasticity" recognizes changeability as the most basic, qualifying property.

A major motivation of the present article is to address a popular custom within neuroscience, mental health research, and popular science to unconditionally equate (if not conflate) markers of neuroplasticity with plasticity proper-and positive mental health change. In the late nineteenth century, inspired by recent theories and discoveries concerning neuronal and synaptic communication, the psychologist William James, and neuroscientists, Santiago Ramón y Cajal, Eugenio Tanzi, and Ernesto Lugaro, each of them referred to the notion of brain "plasticity." All of them, and most notably James, speculated that habit formation and learning must relate to plastic changes in the brain.

Here we argue that these earliest known uses of "plasticity" in neuroscience set the stage for a semantic divergence from plasticity proper. The understandable logic of James and others was that a change in mind or behavior must correspond with a change in brain function or anatomy, and any such changes must require a degree of plasticity in both mind and brain. This meaning of neuroplasticity-as an experience, or stimulus-induced brain change-has remained somewhat consistent over the years. For example, neuroplasticity is commonly defined as the ability of some property of brain function or anatomy to change in response to a stimulus. However, in the present study, we spotlight an issue with this meaning; namely, that it relates to a neuronal event or process that can, paradoxically, predispose the mind and behavior toward the inverse of plasticity proper, namely, "canalization".

First introduced as a philosophical image by Henri Bergson and formalized in developmental biology by Conrad Hal Waddington, canalization describes the process of reducing phenotypic variability (the capacity to vary in response to genetic events or environmental pressures), whether within an organism across its development or between generations via evolutionary mechanisms. In an evolutionary context, canalization refers to the ability of a population of organisms to produce a consistent phenotype by following what Waddington called a chreode: a canalized developmental path. A canalized phenotype remains robust even in the face of environmental or genetic pressures or perturbations. Canalized phenotypes are resistant to change and variation. Plastic phenotypes exhibit the opposite tendency; they vary easily and are sensitive to influence and change.

Several types of neuroplasticity have been named and defined, with varying degrees of specificity and granularity. The following non-exhaustive list includes overlapping constructs that should not necessarily be viewed as independent: functional plasticity, Hebbian plasticity, long-term potentiation and long-term depression, activity-dependent plasticity, spike-timing dependent plasticity, anatomical or structural plasticity, structural remodeling, synaptic plasticity, homosynaptic and heterosynaptic plasticity, spinogenesis, dendritogenesis, neurogenesis, synaptic pruning, metaplasticity, homeostatic plasticity, and synaptic scaling.

Here we spotlight a paradox whereby many popular biomarkers of increased neuroplasticity index part of a (learning) process that favors phenotypic canalization ahead of plasticity. Many of the best-known and studied markers of neuroplasticity are also, more specifically, markers of synaptic plasticity. Focusing on maladaptive (associative) learning and canalization, two stark examples within psychiatry include the pathogenesis of chronic pain and substance use disorders, where pain and reward circuits become sensitized, reinforced, and ultimately canalized in a way that comes to define the disorders, including their stubborn resistance to treatment. There are many examples-like these-of associative learning and related neuroplastic mechanisms driving the development of phenotypic canalization undergirding symptoms of mental illness.

Long-term potentiation and other related Hebbian mechanisms are likely to be involved in the acquisition and entrenchment of many psychopathological habits. Granted, engagement of such mechanisms will not unconditionally strengthen maladaptive phenotypes; for example, LTP (and LTD) could support the learning of new, healthier, more adaptive associations and behaviors, such as can occur with psychotherapy. Still, despite such exceptions, the paradox remains-namely, that activating synaptic and neural plasticity markers (such as LTP) is associated with mechanisms that favor the canalization of maladaptive phenotypes defining mental illnesses. If, as we argue here, and elsewhere, common neuroplasticity markers can index the acquisition of maladaptive phenotypes, then markers and mechanisms of extinction learning (Bouton, twenty oh four) or "unlearning" may be more relevant and conducive to PP (Du et al., twenty twenty-three; Sottile and Vida, twenty twenty-two; Catlow et al., twenty thirteen; Schindler et al., nineteen eighty-six) than the classical markers and mechanisms of Hebbian learning, such as LTP.

In this regard, we note Cajal's early references to early life plasticity and regeneration; phenomena that are more naturally consistent with PP. Also better aligned with PP is "restorative plasticity". Perhaps the divergence between neuroplasticity and PP that began early on within the neurosciences happened because of the appeal of measuring experimentally induced changes in brain function or anatomy, rather than maintaining a respect for the primacy of phenotypic changeability itself. Put succinctly, brain change was privileged over changeability. Another related reason for the divergence is the focus that basic and preclinical neuroscience has placed on brain, biology, and behavior-and how each relates to learning, in particular. Accordingly, scientists have tended to privilege laboratory-friendly assays of neuroplastic brain changes (such as LTP-related synaptic changes) and (animal) behavioral change over PP.

It feels important-if not essential-to understand whether any given increase in neuroplasticity, in a specific context, carries the risk of driving a paradoxical narrowing of a phenotype's capacity for change. If such a phenomenon were to occur, we might refer to it as a "paradoxically canalizing plasticity". Here we highlight that Hebbian mechanisms of acquisition and reinforcement learning are likely to be operative in PCP. We also consider how stress-induced neural and synaptic atrophy (e.g., of the association cortex and hippocampus) may contribute. Excepting (excessive) synaptic pruning, it is not typical to think of atrophy as a driver of canalization. Yet, it is likely to play a role in combination with classical mechanisms of acquisition learning, such as LTP.

Introduced in the late nineteen forties and inspired by the work of the psychologist Donald Hebb, the "Hebbian postulate" refers to the strengthening (or weakening) of selective neural associations (such as pre- and post-synaptic weights) in response to stimuli. LTP and LTD are two especially well-researched neurobiological mechanisms via which synaptic components (such as post-synaptic density composition and/or neurotransmitter release) are known to change in ways that facilitate the learning of specific associations. Hebbian mechanisms are core to associative learning, especially acquisition, operant and reinforcement learning, and classical conditioning. In and of themselves, these associative mechanisms are largely agnostic to mental health outcomes, but they are fundamental to learning and memory. However, there is evidence that the same basic mechanisms also play a role in the pathogenesis of many mental illnesses, as argued before. In such scenarios, PCP becomes centrally relevant, as does the notion of maladaptive learning, including maladaptive avoidance, and learned helplessness.

In what follows, we critique some problematic conventions within neuroscience, namely: (one) using "plasticity" as shorthand for what should properly be called "neuroplasticity"; (two) unconditionally treating the activation of neuroplasticity markers or mechanisms, such as LTP or synaptogenesis, as examples of "plasticity" when the relevant mechanisms tend to favor the opposite outcome, that is, phenotypic canalization (whether pathological or otherwise); and (three) unconditionally inferring mental-health benefits from the activation of any marker of neuroplasticity. Regarding this third point, the construct of "maladaptive plasticity" is relevant. Previous studies have established links between maladaptive plasticity and a wide range of conditions, such as tinnitus, pain, chronic stress, and substance use disorders.

More constructively, we identify a small number of candidate neural biomarkers for the induction of PP, where phenotypic plasticity is maintained. In this regard, phenotypes of core interest include psychological states and traits directly relevant to mental health. We refer to candidate biomarkers of PP processes as "mediational and recalibrative plasticity" markers. Unlike PCP, MR-P markers index processes that support phenotypic plasticity during and after a plasticity window. Somewhat related constructs include "plastic compensation"-that is, adaptive plasticity overcoming maladaptive plasticity, "allostatic recalibration", early-life plasticity or critical periods, "restorative plasticity", and (broadly speaking) "homeostatic plasticity". However, none of these fully capture the phenomenon we wish to describe, hence our introduction of MR-P. Table one highlights definitions of key terms, and a more comprehensive working glossary and taxonomy of key terms is provided as Supplementary Table S one.